Science Gateway Advanced Support Activities in PTI

1. Science Gateway Advanced Support Activities in PTI Marlon Pierce Indiana University

3. OGCE Gateway Tool Adaption & Reuse LEAD LEAD Experiment Builder, XRegistry Interface GFac, XBaya, XRegistry, FTR Eventing System GridChem XBaya GridChem Ultrascan OGCE Re-engineer, Generalize, Build, Test and Release GFac, Eventing System Resource Discovery Service BioVLab OVP/ RST/ MIG TeraGrid User Portal XBaya, GFac ODI GPIR, File Browser Workflow Suite, Gadget Container OGCE Team Bio Drug Screen Swarm->GFac Gadget Container, GTLab, Javascript Cog, XRegistry Interface, Experiment Builder, Axis2Gfac, Axis2 Eventing System, Resource Prediction Service, Swarm EST Pipeline Swarm->GFac Future Grid GFac, Xbaya, … 3 3

4. Gateway Hosting Service • AllocatableTeraGrid Resource providing Virtual Machine hosting of Science Gateways. • This has been a valuable resource for our group. • We should look for ways to expand its usage, such as supporting data collections Courtesy: Mike Lowe, Dave Hancock.

5. BioDrugScreen Portal Support: Josh Rosen and ArchitKulshrestha Collaboration: SamyMeroueh, IUPUI

6. BioDrugScreen • A computational drug discovery resource • Contains millions of pre-docked and pre-scored complexes between thousands of targets from the human proteome and thousands of drug-like small molecules • Allows drug researchers to develop their own scoring functions of calculating how well a drug will interact with a protein.

7. Small drug-like molecules from NCI diversity set are docked into 205 proteasome protein target. Orange area is an identified target area of the protein. Visualization uses JMol. Docking done with Amber on the TeraGrid. Proteins are obtained from the PDB. Samy maintains his own database of protein structures and small molecules.

8. Support Provided • Proteins that have not had their docking and scores calculated need to have these run on the Teragrid. • An Web interface needed to be created so users can submit their own jobs. • We developed the interface between the site’s submissions and the Teragrid using the Swarm Service. • Also prototyping GFac support • Using Flash and javascript, we developed an improved data presentation for the ranking page

9. UltraScan Gateway Staff: Raminder Singh Collaborators: BorriesDemeler and Emre Brookes, UTHSCSA

10. UltraScan Science Gateway • A biophysics gateway for investigating properties and structure-function relationships of biological macromolecules, nanoparticles, polymers and colloids that are implicated in many diseases, including cancer. • High-resolution analysis and modeling of hydrodynamic data from an analytical ultracentrifuge. • TeraGrid as a backup spill over resources but still is one of the heaviest user consuming 1.75 million SU’s in 6 months.

11. UltraScan Advanced Support • Porting to new architectures and parallel performance enhancements. • New workflow implementations, new grid computing and grid middleware support: • Reliability problems with WSGram • Missing job status • Only supports Gram4, needs porting to other middleware • Issues with data movement. • Need Fault tolerance at all levels. • Users decide resources manually, need automated scheduling. Current Architecture

12. UltraScan OGCE Integration • Enhance the perl job submission daemon and monitoring with OGCE GFacservice. • Implement and iteratively enhance fault tolerance. • Port to community account usage withGridshibauditing support. • Support UNICORE to run jobs on other European and Australian resources.

13. GridChem Support: Suresh Marru, Raminder Singh Collaborators: SudhakarPamidighantam, NCSA

14. GridChem Science Gateway • A chemistry/material Science Gateway for running computational chemistry codes, workflows, and parameter sweeps. • Integrates molecular science applications and tools for community use. • 400+ users heavily using TeraGrid. One of the consistentTop 5 TeraGrid Gateway users. • Supports all popular Chemistry applications including Gaussian, GAMESS, NWChem, QMCPack, Amber and MolPro, CHARMM

15. GridChem Advanced Support • GridChem supports single application executions • Advanced support request for supporting workflows • Improved Fault Tolerance

16. GridChem OGCE Integration • OGCE workflow tools wrapped Gaussian & CHARMM chemistry applications • Coupled Butane workflow using Gaussian & CHARMM Integration • 100 member Gaussian parametric sweeps • Integration with Pegasus workflow tools • Ye Fan, Master’s student

17. GridChem Using OGCE Tools Initial Structure Optimized Structure GridChem using OGCE Workflow Tools to construct and execute CHARMM and Gaussian Molecular chemistry Models

18. Future Grid User Portal Support: SiddMaini, ArchitKulshrestha

19. Future Grid User Portal • Our strategy is to build all components as Google Gadgets that interact with REST Services • Can live in iGoogle as well as containers like Drupal. • Take advantage of OpenID and OAuth • Initial target gadgets: Knowledge Base, Amazon EC2 Clients, Inca Clients • Future Work: services, gadgets, and workflows for managing machine images with Xcat.

20. Future Grid KnowledgeBase (FGKB) • Task: • Develop FGKB Web App • Search KB Documents • Technology used: Adobe Flex / PHP / KB REST API

21. Current Status: Basic Search and Retrieval Live URL: http://tinyurl.com/ykaa9gr

22. EC2 Client User Interface Link: http://tinyurl.com/ylkohj7 • See list of images available • Launch them • Terminate them Next Step: Add more EC2 features, integrate with FutureGrid user database

23. Portal Embedded Gadgets

24. FutureGrid Machine Image Services • FutureGrid will use Xcat to dynamically create and manage clusters from preconfigured images • On both real hardware and Virtual machines. • We are working to capture common XCat tasks as scripts • These scripts can be wrapped as secure services using OGCE’sGFac. • Several tasks can be linked together as workflows visually composed with OGCE’sXBaya. • You can still use Pegasus/Condor as the workflow engine

25. OGCE Software for Science Gateways Software and Architectural Approach

26. Science Gateways Layer Cake User Interfaces Gateway Abstraction Interfaces Web Enabled Desktop Applications Web/Gadget Container Web/Gadget Interfaces Information Services Application Abstractions Application Monitoring User Management Fault Tolerance Gateway Services Provenance & Metadata Management Registry Workflow System Security Auditing & Reporting Resource Middleware Cloud Interfaces SSH & Resource Managers Grid Middleware Compute Resources Local Resources Computational Grids Computational Clouds Color Coding OGCE Gateway Components Complimentary Gateway Components Dependent resource provider components

27. Google Gadget-Based Science Gateways PolarGrid MyOSG LEAD

28. GFac Current & Future Features Apache Axis2 Globus Registry Interface Scheduling Interface Input Handlers Monitoring Interface Campus Resources Output Handlers Fault Tolerance Data Management Abstraction Amazon Eucalyptus Job Management Abstraction Auditing Checkpoint Support Unicore Condor Color Coding Existing Features Planned/Requested Features

29. OGCE Layered Workflow Architecture:Derived from LEAD Workflow System Workflow Interfaces (Design & Definition) XBaya GUI (Composition, Deploying, Steering & Monitoring) Flex/Web Composition Gadget Interface for Input Binding BPEL 2.0 Python Workflow Specification Scufl BPEL 1.0 Java Code Pegasus DAG Workflow Execution & Control Engines Apache ODE Dynamic Enactor Condor DAGMan Jython Interpreter GBPEL Taverna

30. Putting It All Together

31. Software Strategy • Focus on gadget container and tools for running science applications on grids and clouds. • Provide a tool set that can be used in whole or in part. • If you just want GFac, then you can use it without buying an entire framework. • Outsource security, information services, data and metadata, advanced job handling, etc to other providers. • MyProxy, TG IIS, Globus, Condor, XMC Cat, iRods, etc.

32. Packaging, Building, and Testing • All builds are designed to be self contained • Use Apache Maven 2.x • Download includes everything you need. • Portal, Axis Services, and Xregistry all build nightly on the NMI Build and Test facility at UW. • Several Linux platforms, Mac PPC, and Mac X86 • Java 1.5 • Apache JMetertest suite for the portal. • Run against your installation • Automated tests nightly

33. Next Steps • Apache Incubator Project for XBaya, GFac and supporting workflow tools • WIYN ODI instrument pipeline and gateway • Robert Henschel and Scott Michael are leading overall effort. • Suresh and Raminder are working 50% time through early April on technical evaluation of integrating NHPPS software with OGCE • New Collaboration: Craig Mattocks, UNC will build a storm surge forecasting gateway • Broadly similar to LEAD and SCOOP • Archit will be the point of contact • Local gateway tutorial in early April • Sun Kim’s group, UNC Group, CGB • Others welcome • Gadget Container additional applications • OGCE grid gadgets packaged release, SimpleGrid

34. More Information, Acknowledgements • Website: www.collab-ogce.org • Blog/RSS Feed: collab-ogce.blogspot.com • Email: mpierce@cs.indiana.edu, smarru@cs.indiana.edu • Geoffrey Fox, Craig Stewart, and Dennis Gannon have spent years laying the foundation for this work

35. Backup Slides

36. EST Assembly Pipeline Support: ArchitKulshrestha, Chin Hua Kong Collaborator: Qunfeng Dong, UNT

37. Our goal is to provide a Web service-based science portal that can handle the largest mRNA clustering problems. • Computation is outsourced to Grids (TeraGrid) and Clouds (Amazon) • Not provided by in-house clusters. • This is an open service, open architecture approach. • These codes have very different scheduling requirements http://swarm.cgb.indiana.edu

38. EST Assembly Pipeline • OGCE SWARM is used to intelligently submit thousands of jobs to compute resources of various sizes such as workstations and Grid enabled supercomputers. • TeraGrid’sBigRed, Ranger, and Cobalt: PACE, RepeatMasker • Local clusters: CAP3 • Support for Job submission to the Cloud is being developed and will address the need for resources larger (in terms of available memory) than clusters currently available.

39. UltraScan Gateway Staff: Raminder Singh Collaborators: BorriesDemeler and Emre Brookes, UTHSCSA

40. UltraScan Science Gateway • A biophysics gateway for investigating properties and structure-function relationships of biological macromolecules, nanoparticles, polymers and colloids that are implicated in many diseases, including cancer. • High-resolution analysis and modeling of hydrodynamic data from an analytical ultracentrifuge. • TeraGrid as a backup spill over resources but still is one of the heaviest user consuming 1.75 million SU’s in 6 months.

41. UltraScan Advanced Support • Porting to new architectures and parallel performance enhancements. • New workflow implementations, new grid computing and grid middleware support: • Reliability problems with WSGram • Missing job status • Only supports Gram4, needs porting to other middleware • Issues with data movement. • Need Fault tolerance at all levels. • Users decide resources manually, need automated scheduling. Current Architecture

42. UltraScan OGCE Integration • Enhance the perl job submission daemon with OGCE Gfac service. • Enhance socket and email based job monitoring with OGCE Eventing System • Implement and iteratively enhance fault tolerance. • Port to Community account usage with gridshib auditing support. • Support Unicore to run jobs on other European and Australian resources.

43. OGCE based UltraScan development Architecture Manual Process Quarry Gateway Hosting Machine UltraScan Middleware GFac, Eventing System, Fault Tolerance Europe & Australian Grids

44. Future Grid KB Image Manager Experiment Browser INCA Monitor Downloadable Gadgets

45. Deployment (in future) iGoogle Gadgets Image Management FG INCA MONITOR Experiment Management iPhone Application

46. JMeter test suite

47. Why Gadgets? • We have redesigned many OGCE components to work as gadgets. • Fugang Wang’s Cyberaide JavaScript gives us an API • Framework and language independent • Client-side HTML, CSS, and JavaScript, not server-side Java • Integration and content under user’s control, not portal administrator’s • Can be integrated into iGoogle and similar containers. • 140,000 published gadgets • Joomla, Drupal, Liferay, etc • We can potentially provide HUBzero gadgets

48. OGCE based UltraScan development Architecture Manual Process Quarry Gateway Hosting Machine UltraScan Middleware GFac and supporting services Europe & Australian Grids

49. BioDrugScreen Next Steps • We want to expand the user generated function process, including the ability for a user to save a function and have multiple functions • Interaction between users will be enhanced, allowing them to share their functions and findings.